Bekker, who previously founded Voca and sold it to Snap for $100 million, has a background in voice recognition, neural networks, and deep learning. After leaving Snap, he envisioned a new generation of AI agents using video and voice for more humanlike interactions. eSelf leverages fine-tuned models like Meta's Llama rather than building its own large language models from scratch, offering an alternative to existing chatbot solutions with potentially lower latency. The startup is part of a growing field of companies exploring the "agentic" opportunity, aiming to deliver more believable AI responses through advanced generative AI technologies.
Key takeaways:
```html
- eSelf, a startup co-founded by Dr. Alan Bekker and Eylon Shoshan, has emerged from stealth with $4.5 million in seed funding to develop real-time, video-based conversational AI agents.
- The platform focuses on verticals such as education, sales, financial services, real estate, and health and wellness, offering a "no-code" creation studio for building AI agents.
- eSelf's agents boast a response time of under 1.7 seconds, claiming to be faster and more lifelike than competitors like OpenAI.
- The startup leverages fine-tuned Meta-created Llama models instead of building its own LLMs, allowing customers to switch from existing chatbots with potentially longer latency.